Summary On 12 April 2005, a Lockheed L382G Hercules (registration C-GHPW, serial number 4799) owned and operated by Bradley Air Services Ltd. (First Air) and designated as flight FAB702, departed High Lake, Nunavut, for Yellowknife, Northwest Territories, with four crew members on board. At 1139 mountain daylight time (MDT), approximately 10minutes after departure, as the aircraft was climbing through 18000feet, the crew heard a bang from the cargo area. When they examined the cargo compartment, they heard the sound of air escaping from the left side of the compartment and discovered a crack estimated to be 24inches long and approximately one-half inch wide in the left wing-to-fuselage attach angle (drag angle). When the flight crew learned there was a major structural failure, the aircraft was levelled off at flight level (FL)230 and depressurized. Speed was reduced to 180KIAS (knots indicated airspeed), an emergency was declared, and all crew members went on oxygen. The aircraft was level for about five minutes then descended to FL220 for the direction of flight, and remained level for about 35 minutes. The flight crew later descended to 10000feet to ensure they were well supplied with oxygen. By this time the crack was no longer visible. On nearing Yellowknife, the aircraft was slowed to 140KIAS (rather than 170KIAS) when the landing gear was lowered. The aircraft landed safely at 1312MDT with the flap retracted. Aircraft rescue and firefighting crews and equipment were standing by. There were no injuries. Ce rapport est galement disponible en franais. Other Factual Information The crew was certified and qualified for the flight in accordance with existing regulations and very experienced in Lockheed L382G Hercules operations. After the crack was discovered, the crew had several options, including returning to High Lake, Nunavut, or diverting to an alternate airport. The runway at High Lake was a 5150-foot-long ice strip that was unsuitable for a precautionary zero-flap landing, and all the other available alternates were gravel airstrips where aircraft rescue and firefighting (ARFF) services were unavailable. The decision to continue to Yellowknife, Northwest Territories, was predicated on the availability of ARFF and a 7500-foot-longrunway. Visual meteorological conditions, with generally clear skies and light winds, prevailed for the entire flight. Weather was not considered a factor in the occurrence. After the aircraft landed, at 1312,1 the crack was inspected; it had closed up and was not visible. The aircraft was subsequently flown unpressurized to a repair facility in Edmonton, Alberta. It was determined that the crack had propagated to 33inches in length in the radius of the left attach angle, PN390624-7, near fuselage station (FS)577, butt line (BL)61. Records indicate the attach angle had accumulated 35789hours time in service and had undergone 31205pressure cycles. The one-piece left and right wing-to-fuselage attach angles2 secure the fuselage sidewall to the lower skin of the centre wing box beam in the Dashed lines indicate wing attach angle. aircraft. The attach angles are installed on the upper corners of the cabin, from forward of the centre wing at FS477 to aft of the centre wing at FS617 (see Figure1). They provide a pressurization seal at the centre wing-to-fuselage juncture and support the fuselage sidewall against pressurization blow out (see Photo1). Primary wing-to-fuselage structural loads are carried by vertical beams extending down the side of the fuselage, below the centre wing. To date, attach angle cracking has never contributed to the loss of an aircraft due to catastrophic structural failure. The attach angles installed on L382/C-130 aircraft manufactured prior to serial number5306 are machined from a 7075-T6 aluminum extrusion. To increase the bending strength of the attach angles in aircraft serial numbers4383 to 5305, strap doublers are installed on the vertical and horizontal legs of the attach angles from FS477 through FS617. Many Lockheed Hercules aircraft, including the L382G stretched versions, have additional thick step doublers installed aft of FS588, below the rear spar. The occurrence aircraft had both strap doublers and step doublers. As of serial number 5306, aircraft are fitted with improved attach angles manufactured from 7050aluminum. These attach angles are thicker and do not require reinforcing straps to be fitted to the legs. However, they are not dimensionally compatible with earlier aircraft due to the location of the fastener holes in the lower wing skin. The manufacturer's original recommendations for inspecting and servicing cracks in the wing-to-fuselage attach angles allowed stop drilling and repair of cracks. Lockheed Service Bulletin (SB) 382-53-61/82-752 Basic and Revision1 specified the overriding requirements for inspection of wing-to-fuselage attach angles in commercial and military L382/C-130 aircraft and included an X-ray inspection from FS517 to FS617. However, due to the difficulty in placing the x-ray film behind the attach angle, many operators were not able to perform this inspection with confident results. Visual inspection was also ineffective due to the position and thickness of the reinforcing strap and step doublers and to the application of sealant in aircraft serial numbers4383 to5305. Numerous other inspection methods were tried. However, a suitable method short of extensive disassembly of the subject area could not be found. A review of the in-service cracking data by the manufacturer indicated that the areas adjacent to FS497 and FS577 were both common origins for cracks. Reportedly, all attach angles that had cracking at FS577 also had cracking at FS497. Relatively small cracks at FS497 could be detected by existing x-ray inspection, but the area near FS577 could not be easily examined by any usual inspection method. Cracking at FS497 will not result in pressure loss until the crack becomes large enough to extend aft of FS517. SB 382-53-61/82-752, Revision 2, dated 28February2003, specified the manufacturer's current requirements for inspection of the wing-to-fuselage attach angles in commercial and military L382/C-130 aircraft. The requirement to inspect the areas between FS517 and 617was removed. Instead, Revision2 recommended replacing the attach angle once cracks are detected at FS497 by x-ray examination. The left attach angle on the occurrence aircraft had already been repaired at FS497 on 08September1987 in accordance with Lockheed Aircraft Corporation drawing SK382-57-093 at 14498hours and 13282cycles. Revision2 did not address replacing previously repaired attach angles. The left wing-to-fuselage attach angle of the occurrence aircraft was removed during repair and shipped to the TSB Engineering Branch to determine the mode of failure. Examination of the crack at the forward end of the attach angle, near FS497, determined that the crack had resulted from the overload extension of some 10or 11small fatigue-generated pre-cracks. The forward crack had been stop drilled and re-enforced during the repair that had been accomplished in1987, and no further growth of the crack had occurred in the 18years since that repair. The crack in the radius at the aft end of the attach angle, near FS577, resulted from the initiation of three regions of fatigue cracking that had extended as a result of an Arrows indicate crack in attach angle. overstress to form a single crack measuring 33inches in length (see Photo2). Measurements of crack growth rates suggested that the crack was growing slightly every time the fuselage was pressurized and had been active over a number of years. No indication of stress corrosion was found in the crack. In communications with Lockheed representatives during this investigation, they stated that stress corrosion cracking had been the mechanism of crack initiation in all previous attach angle failures. The number of failures was not made known to TSBinvestigators. The occurrence aircraft had undergone a Ccheck at Aero Aviation in Calgary, Alberta, during January to March2005, at 35708:31hours and 31147cycles. First Air was authorized to perform line maintenance, and Aand Bchecks on the L-382G. Transport Canada issued a one-time authorization to First Air to allow the company to perform the final maintenance release following the Ccheck. Certification of the inspection and the associated repairs was completed by a First Air representative, and the aircraft was returned to service on 27March2005. No cracking was identified during the examination of the forward repaired section of the left attach angle during the C check. Cracking was discovered in the right attach angle at FS497, and the right attach angle was repaired in accordance with Aerotech Structures Inc. Drawing #R05-002-01, Rev.1R. This was an FAA (United States Federal Aviation Administration) Designated Engineer Representative (DER) FAA8110-3 approval. Although Lockheed Service Bulletin 382-53-61/82-752, Revision2 recommends replacing an attach angle whenever cracking is detected, a DER approval can override an SB. The repair at FS497 restored the right attach angle to original strength at that location. However, the repair approval did not include a continuing maintenance program covering inspection for cracks at FS577. Weight and balance was not a factor in this occurrence. However, the investigation identified a concern relating to the use of a specific electronic, low-profile, platform scale system for weighing aircraft. Ramp or platform type scale systems are commonly used to weigh large aircraft, one advantage being that an aircraft can be weighed without using jacks. Although the system recently used to weigh the occurrence aircraft had been marketed as an aircraft scale system, it had not been specifically designed for that purpose. Electronic scale systems can be assembled from a variety of digital indicator and electronic platform combinations. Aero Aviation had purchased the portable electronic scale system in March 2005. The system, supplied by A&A ScalesLLC, consisted of six 40000-pound capacity Massload Technologies Ultra Slim Weigh Pad electronic, low-profile, platform scales and two Tara Systems ModelTR-1-NK digital weight indicators. The weigh pads were lightweight and portable and were designed primarily for weighing transport trucks. They have been used only sporadically in the aviation industry. The weigh pads measured 28.5inches by 16.5inches by 0.7inches thick and contained no moving parts. Twenty transducers were embedded into the bottom of each light-weight aluminum platform. The weigh pads had to be level and well supported by the underlying floor structure when in use. The two digital weight indicators provided for single-point calibration. Scale accuracy is known to increase when the systems use digital indicators that provide multiple point calibration over numerous load points and that are calibrated to each size of tire footprint. The system was first used to weigh the occurrence aircraft on 22March2005. One digital indicator was connected to platforms placed under the nose wheels and the other digital indicator was connected to platforms placed under the main wheels. The aircraft was weighed twice using the new scale system, with consistent results, and the new basic empty weight was recorded as 71039pounds. This corresponded to a reduction of 3943pounds compared to the aircraft basic empty weight of 74982pounds from the previous weight and balance report, which had been prepared in October of2003. An estimated 222pounds of equipment had been added to the aircraft since October2003; however, review of the October2003 weight and balance report indicated a large, unexplained increase in the aircraft empty weight over the previous weight and balance report. Based on the unexplained weight increase in the October2003 weight and balance report, and on the fact that the aircraft had been weighed twice with consistent results, the aircraft was returned to service with the basic empty weight recorded as 71039pounds. The discrepancy was re-examined approximately 75hours after the aircraft was returned to service, and the October2003 weight and balance report was temporarily reinstated. As a precaution, an overweight landing inspection of the aircraft was performed on 11April2005. The aircraft was re-weighed on 13September2005, and the basic empty weight was determined to be 75477pounds, which was 4438pounds heavier than the weight recorded on 22March2005. The electronic scale system had been calibrated on 11March2005 before it was delivered to Aero Aviation. The calibration plate used was only slightly smaller than the platform surface. After the aircraft had been weighed, First Air requested that the system be re-calibrated. The system was re-calibrated on 05April2005 using a smaller calibration plate. The revised calibration data identified that the indicators were displaying weights between 240and 360pounds lower than actual, per pad, with an applied load of 5000pounds, and between 760and 1260pounds lower than actual, per pad, with an applied load of 20000pounds. Aero Aviation returned the scale system to the vendor for refund following re-calibration, and the vendor no longer markets scale systems for weighing aircraft. Aero Aviation replaced the system with single-point, jack load cell type equipment.